Physicl raises $10M to build realistic 3D simulations for training robots

A new startup named Physicl has secured $10 million in seed funding to help robot builders solve training challenges. Inovia Capital led the funding round. Founder and CEO Alex de Vigan previously spent a decade building a platform that created 3D product models for e-commerce. When generative artificial intelligence changed that market, he realized his team’s technology could be repurposed to help robots understand physical space.

Training robots with synthetic data

Researchers trying to make robots handle everyday objects often struggle with training limitations. Training models on real video can be inefficient, and gathering data through remote operations is expensive. Digital simulations offer a promising alternative, but making those virtual worlds obey real-world physics is difficult. Physicl aims to use its 3D rendering pipeline to turn unstructured real-world information into accurate physical representations.

The Physicl Studio platform

The company is launching a product called Studio. It includes thousands of 3D environments and assets to help users generate synthetic data for specific problems. For instance, the software can help train a robot to find an object on a cluttered counter. Early customers like world model builder SpAItial are already using these tools to improve the quality of their 3D simulations derived from flat images.

• Physicl raised $10 million in seed funding led by Inovia Capital.
• The company launched Studio with thousands of 3D environments and assets.
• Early partners use the platform to generate synthetic training data for robots.